[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: [ldapext] dc+idn



Kurt D. Zeilenga wrote:
At 03:44 PM 11/10/2004, Leif Johansson wrote:

Kurt D. Zeilenga wrote:

At 02:31 PM 11/8/2004, Leif Johansson wrote:


Could someone summarize the state of affairs with "internationalized
dc"? It seems to me that there is an urgent need looming here.

First, here are some earlier starts on this: http://www.watersprings.org/pub/id/draft-zeilenga-ldap-idn-04.txt http://www.watersprings.org/pub/id/draft-hall-ldap-idn-00.txt However, at present, I'm thinking that internationalization of domain components (and like things) is better handled above the directory.... otherwise internationalized and non-internationalized clients will not interoperate properly.

Well that may be so but It occured to me the other day that one way to allow the directory to expose idn's to the user is to create a transfer option (;idna-decode or something to that effect) which would return dc and associatedDomain in the ACE-decoded form. A quick read of the application requirements from RFC3490 seems to indicate that a transfer option might be constructed that would fullfill these requirements.


The problem with this approach is that until the transfer
encoding is widely implemented and deployed, clients could
not rely on its presence.  With the above-the-directory
approach, internationalized clients can be developed and
deployed today need for internationalized support in the
servers.


Right, but this is no worse than what the client can expect today! Today
the client will se ACE-encoded values in dc and associatedDomain. Those
are perfectly legal domain names. The transfer option simply allows
clients and servers who implement this option to retrieve the
ACE-decoded values. This would naturally depend on feature-negotiation
(via say the root-DSE).

	MVH leifj


_______________________________________________ Ldapext mailing list Ldapext@ietf.org https://www1.ietf.org/mailman/listinfo/ldapext